Finding experiments

To use incense we first have to instantiate an experiment loader that will enable us to query the database for specific runs.

targets_type iteration autoencoder_type batch_size artifacts
exp_id
34 Mnist False Over_dim_tied 256 {'history_autoencoder': Artifact(name=history_...
35 Mnist False Over_dim_tied 128 {'history_autoencoder': Artifact(name=history_...
36 Mnist False Over_dim_tied 64 {'history_autoencoder': Artifact(name=history_...
37 Mnist False Over_dim_tied 32 {'history_autoencoder': Artifact(name=history_...
38 10_Targets False Over_dim_tied 256 {'history_autoencoder': Artifact(name=history_...
39 10_Targets False Over_dim_tied 128 {'history_autoencoder': Artifact(name=history_...
40 10_Targets False Over_dim_tied 64 {'history_autoencoder': Artifact(name=history_...
41 10_Targets False Over_dim_tied 32 {'history_autoencoder': Artifact(name=history_...
targets_type iteration autoencoder_type batch_size artifacts sort
exp_id
38 10_Targets False Over_dim_tied 256 {'history_autoencoder': Artifact(name=history_... 0
39 10_Targets False Over_dim_tied 128 {'history_autoencoder': Artifact(name=history_... 1
40 10_Targets False Over_dim_tied 64 {'history_autoencoder': Artifact(name=history_... 2
41 10_Targets False Over_dim_tied 32 {'history_autoencoder': Artifact(name=history_... 3
34 Mnist False Over_dim_tied 256 {'history_autoencoder': Artifact(name=history_... 4
35 Mnist False Over_dim_tied 128 {'history_autoencoder': Artifact(name=history_... 5
36 Mnist False Over_dim_tied 64 {'history_autoencoder': Artifact(name=history_... 6
37 Mnist False Over_dim_tied 32 {'history_autoencoder': Artifact(name=history_... 7

Red best overall, and also best of subset. Bes means for accuracy max, rest min. Green best of subset.

predictions_df_0
Accuracy over iterations evaluations_feature_classifier
Over_dim_tied 256 10_Targets Over_dim_tied 128 10_Targets Over_dim_tied 64 10_Targets Over_dim_tied 32 10_Targets Over_dim_tied 256 Mnist Over_dim_tied 128 Mnist Over_dim_tied 64 Mnist Over_dim_tied 32 Mnist
0 0.9838 0.981 0.966 0.9742 0.9777 0.9777 0.9784 0.9798
1 0.9823 0.9764 0.939 0.9689 0.9764 0.9742 0.9782 0.9782
2 0.9821 0.9759 0.8388 0.9691 0.9738 0.9719 0.9747 0.9698
3 0.9821 0.9758 0.8187 0.9689 0.9691 0.9649 0.9691 0.9546
4 0.9821 0.9758 0.7619 0.9689 0.9628 0.9585 0.962 0.9336
5 0.9821 0.9758 0.7616 0.9689 0.9544 0.9484 0.9515 0.9058
6 0.9821 0.9758 0.7615 0.9689 0.9416 0.9347 0.9371 0.8746
7 0.9821 0.9758 0.7614 0.9689 0.9274 0.9199 0.9211 0.841
Loss over iterations autoencoder
Over_dim_tied 256 10_Targets Over_dim_tied 128 10_Targets Over_dim_tied 64 10_Targets Over_dim_tied 32 10_Targets Over_dim_tied 256 Mnist Over_dim_tied 128 Mnist Over_dim_tied 64 Mnist Over_dim_tied 32 Mnist
0 0.402843 0.406489 0.395777 0.400077 0.0142273 0.0135879 0.0154778 0.0284019
1 0.407147 0.800692 0.408306 0.406918 0.0220556 0.0219748 0.0252352 0.0442787
2 0.407382 2.44856e+10 0.427664 0.407827 0.0327176 12804.8 0.0380625 0.0643806
3 0.4074 1.6029e+21 0.430916 0.407958 0.0459869 12805 19018.1 0.0870223
4 0.407401 inf 0.442703 0.407966 164.111 12805 1.53874e+19 0.112242
5 0.407401 inf 0.446805 0.407966 3161.82 25606.4 inf 1.19025e+12
6 0.407401 inf 0.448094 0.407967 6155.18 114887 inf 4.31161e+27
7 0.407401 inf 0.448288 0.407967 15973.5 202682 inf inf
MAE over iterations autoencoder
Over_dim_tied 256 10_Targets Over_dim_tied 128 10_Targets Over_dim_tied 64 10_Targets Over_dim_tied 32 10_Targets Over_dim_tied 256 Mnist Over_dim_tied 128 Mnist Over_dim_tied 64 Mnist Over_dim_tied 32 Mnist
0 0.26689 0.270193 0.273218 0.268186 0.0414262 0.0394306 0.0430273 0.0627057
1 0.267821 0.276881 0.278261 0.269768 0.0505998 0.0492101 0.0539312 0.0779905
2 0.267859 1364.28 0.286771 0.269981 0.0606934 1.13672 0.0655467 0.0939986
3 0.267863 3.48994e+08 0.289726 0.270016 0.0711264 1.14721 1.40036 0.109801
4 0.267863 8.92929e+13 0.294124 0.270018 0.202172 1.15747 3.76367e+07 0.125302
5 0.267863 2.28463e+19 0.295313 0.270018 2.11776 2.24865 1.07062e+15 9855.67
6 0.267863 5.84539e+24 0.295689 0.270018 4.12205 10.1626 3.04553e+22 5.93163e+11
7 0.267863 1.49559e+30 0.295747 0.270018 10.5071 17.2467 8.66343e+29 3.57005e+19 
predictions_df_10
Accuracy over iterations evaluations_feature_classifier
Over_dim_tied 256 10_Targets Over_dim_tied 128 10_Targets Over_dim_tied 64 10_Targets Over_dim_tied 32 10_Targets Over_dim_tied 256 Mnist Over_dim_tied 128 Mnist Over_dim_tied 64 Mnist Over_dim_tied 32 Mnist
0 0.9783 0.9749 0.9585 0.9671 0.9527 0.9529 0.9514 0.9637
1 0.9744 0.9693 0.9306 0.961 0.9626 0.9598 0.9633 0.9674
2 0.9742 0.9685 0.8275 0.9602 0.9604 0.9561 0.9615 0.9584
3 0.9741 0.9685 0.8081 0.96 0.9554 0.9489 0.9571 0.9379
4 0.9741 0.9685 0.756 0.96 0.9472 0.9372 0.946 0.9139
5 0.9741 0.9685 0.7559 0.96 0.9362 0.9239 0.9296 0.8838
6 0.9741 0.9685 0.7557 0.96 0.9183 0.9087 0.9121 0.8503
7 0.9741 0.9685 0.7557 0.96 0.9004 0.8903 0.89 0.8142
Loss over iterations autoencoder
Over_dim_tied 256 10_Targets Over_dim_tied 128 10_Targets Over_dim_tied 64 10_Targets Over_dim_tied 32 10_Targets Over_dim_tied 256 Mnist Over_dim_tied 128 Mnist Over_dim_tied 64 Mnist Over_dim_tied 32 Mnist
0 0.402125 0.405372 0.393313 0.398376 0.0373236 180516 626119 0.043688
1 0.408081 0.413748 0.408405 0.407693 0.042127 204880 5.06601e+20 0.0563758
2 0.408712 0.414783 0.428755 0.409104 0.0533446 204880 inf 0.0752805
3 0.408743 0.414965 0.432285 0.409356 422.336 204880 inf 0.0972759
4 0.408748 0.414991 0.443466 0.409381 1687.51 217685 inf 0.121509
5 0.408748 0.414992 0.447229 0.409387 4429.59 256100 nan 3.38791e+11
6 0.408748 0.414992 0.448405 0.409387 10821.5 288666 nan 1.22725e+27
7 0.408748 0.414992 0.448596 0.409387 18500.3 382646 nan inf
MAE over iterations autoencoder
Over_dim_tied 256 10_Targets Over_dim_tied 128 10_Targets Over_dim_tied 64 10_Targets Over_dim_tied 32 10_Targets Over_dim_tied 256 Mnist Over_dim_tied 128 Mnist Over_dim_tied 64 Mnist Over_dim_tied 32 Mnist
0 0.267178 0.270283 0.273126 0.268244 0.0675787 15.6296 35.387 0.078606
1 0.268489 0.272088 0.278816 0.270371 0.0697121 17.3004 1.00472e+09 0.0880597
2 0.268611 0.272339 0.287446 0.270712 0.0768396 17.3076 2.85805e+16 0.101744
3 0.268613 0.272385 0.290337 0.27079 0.362277 17.3163 8.13011e+23 0.116358
4 0.268614 0.272392 0.294468 0.270797 1.17568 18.4025 2.31272e+31 0.1311
5 0.268614 0.272392 0.295559 0.270798 2.94163 21.6426 nan 5262.17
6 0.268614 0.272392 0.295901 0.270798 7.11081 24.5996 nan 3.16698e+11
7 0.268614 0.272392 0.295957 0.270798 12.0548 32.3657 nan 1.9061e+19 
predictions_df_20
Accuracy over iterations evaluations_feature_classifier
Over_dim_tied 256 10_Targets Over_dim_tied 128 10_Targets Over_dim_tied 64 10_Targets Over_dim_tied 32 10_Targets Over_dim_tied 256 Mnist Over_dim_tied 128 Mnist Over_dim_tied 64 Mnist Over_dim_tied 32 Mnist
0 0.9661 0.9636 0.9461 0.9594 0.9201 0.9149 0.9217 0.9304
1 0.9557 0.9583 0.914 0.945 0.9369 0.9295 0.9425 0.942
2 0.9551 0.9579 0.8126 0.9439 0.9346 0.9257 0.9425 0.9328
3 0.9551 0.9578 0.7909 0.9438 0.9261 0.9156 0.9351 0.9137
4 0.9551 0.9578 0.7443 0.9436 0.9119 0.8986 0.9222 0.8827
5 0.9551 0.9578 0.7437 0.9436 0.8959 0.8813 0.8962 0.8488
6 0.9551 0.9578 0.7437 0.9436 0.8753 0.8604 0.8758 0.8135
7 0.9551 0.9578 0.7437 0.9436 0.8522 0.8391 0.8511 0.7701
Loss over iterations autoencoder
Over_dim_tied 256 10_Targets Over_dim_tied 128 10_Targets Over_dim_tied 64 10_Targets Over_dim_tied 32 10_Targets Over_dim_tied 256 Mnist Over_dim_tied 128 Mnist Over_dim_tied 64 Mnist Over_dim_tied 32 Mnist
0 0.40102 0.403685 0.390298 0.397119 0.0590926 883936 3.31396e+06 0.0599732
1 0.410223 0.415205 0.408881 0.410071 0.0627645 1.0116e+06 2.68146e+21 0.0705858
2 0.411363 0.416537 0.429943 0.412499 211.014 1.0116e+06 inf 0.0895225
3 0.411388 0.41663 0.434311 0.412865 1131.21 1.02434e+06 inf 6.48896e+11
4 0.411388 0.416633 0.444758 0.413032 4997.33 1.0244e+06 inf 2.35058e+27
5 0.411388 0.416633 0.448318 0.413036 12822.6 1.075e+06 nan inf
6 0.411388 0.416633 0.4494 0.413036 19983.4 1.15187e+06 nan inf
7 0.411388 0.416633 0.449552 0.413036 31805.6 1.2014e+06 nan inf
MAE over iterations autoencoder
Over_dim_tied 256 10_Targets Over_dim_tied 128 10_Targets Over_dim_tied 64 10_Targets Over_dim_tied 32 10_Targets Over_dim_tied 256 Mnist Over_dim_tied 128 Mnist Over_dim_tied 64 Mnist Over_dim_tied 32 Mnist
0 0.267796 0.270376 0.273176 0.268869 0.0868673 77.4841 126.485 0.0935081
1 0.269855 0.272937 0.279665 0.272002 0.0863927 85.1702 3.5957e+09 0.0993183
2 0.270111 0.273244 0.288405 0.272628 0.227075 85.1759 1.02285e+17 0.111268
3 0.270116 0.27326 0.291378 0.272753 0.858623 86.2577 2.90962e+24 7277.04
4 0.270116 0.273261 0.29518 0.272812 3.39602 86.2688 8.2768e+31 4.37967e+11
5 0.270116 0.273261 0.29622 0.272813 8.38942 90.5591 nan 2.63598e+19
6 0.270116 0.273261 0.296538 0.272813 12.9823 97.0312 nan 1.58651e+27
7 0.270116 0.273261 0.296583 0.272813 20.6104 101.723 nan inf 
predictions_df_30
Accuracy over iterations evaluations_feature_classifier
Over_dim_tied 256 10_Targets Over_dim_tied 128 10_Targets Over_dim_tied 64 10_Targets Over_dim_tied 32 10_Targets Over_dim_tied 256 Mnist Over_dim_tied 128 Mnist Over_dim_tied 64 Mnist Over_dim_tied 32 Mnist
0 0.9459 0.9462 0.9297 0.9427 0.8829 0.8651 0.878 0.8838
1 0.9316 0.9347 0.8946 0.9163 0.9111 0.8875 0.9118 0.9127
2 0.9298 0.933 0.7884 0.9134 0.908 0.8798 0.9119 0.8996
3 0.9297 0.9329 0.765 0.9129 0.8948 0.8657 0.9016 0.8721
4 0.9297 0.9329 0.7283 0.9127 0.8783 0.8454 0.8854 0.8408
5 0.9297 0.9329 0.728 0.9126 0.8573 0.823 0.8549 0.8065
6 0.9297 0.9328 0.7279 0.9126 0.8332 0.7974 0.8295 0.7662
7 0.9297 0.9328 0.7279 0.9126 0.8056 0.7744 0.8026 0.7276
Loss over iterations autoencoder
Over_dim_tied 256 10_Targets Over_dim_tied 128 10_Targets Over_dim_tied 64 10_Targets Over_dim_tied 32 10_Targets Over_dim_tied 256 Mnist Over_dim_tied 128 Mnist Over_dim_tied 64 Mnist Over_dim_tied 32 Mnist
0 0.401364 0.403679 0.387011 0.395785 0.0812379 1.70153e+06 6.24131e+06 0.0780647
1 0.415957 1.63254e+07 0.410203 0.415065 0.0867466 1.81831e+06 5.05017e+21 4.86324e+07
2 0.417589 1.06871e+18 0.433102 0.419083 843.909 1.81831e+06 inf 1.76167e+23
3 0.417732 6.99609e+28 0.438297 0.419722 4594.69 1.83112e+06 inf inf
4 0.417732 inf 0.44764 0.41987 11155.7 1.86742e+06 inf inf
5 0.417732 inf 0.450753 0.419932 17289.7 1.93401e+06 nan inf
6 0.417732 inf 0.451669 0.419934 24303.5 2.05957e+06 nan nan
7 0.417732 inf 0.451814 0.419934 35736.4 2.22487e+06 nan nan
MAE over iterations autoencoder
Over_dim_tied 256 10_Targets Over_dim_tied 128 10_Targets Over_dim_tied 64 10_Targets Over_dim_tied 32 10_Targets Over_dim_tied 256 Mnist Over_dim_tied 128 Mnist Over_dim_tied 64 Mnist Over_dim_tied 32 Mnist
0 0.269936 0.271745 0.273785 0.270188 0.103917 145.311 189.433 0.108527
1 0.273229 35.4951 0.281374 0.275274 0.102642 153.039 5.38588e+09 80.9836
2 0.273623 9.01144e+06 0.290365 0.276372 0.649036 153.044 1.53208e+17 4.86721e+09
3 0.273652 2.30565e+12 0.293345 0.276571 3.07437 154.128 4.35822e+24 2.92941e+17
4 0.273652 5.89917e+17 0.296698 0.276628 7.34871 157.274 1.23975e+32 1.76312e+25
5 0.273652 1.50935e+23 0.297607 0.276648 11.2381 162.963 nan inf
6 0.273652 3.86178e+28 0.297878 0.276649 15.828 173.899 nan nan
7 0.273652 inf 0.297921 0.276649 23.1719 187.924 nan nan 
predictions_df_40
Accuracy over iterations evaluations_feature_classifier
Over_dim_tied 256 10_Targets Over_dim_tied 128 10_Targets Over_dim_tied 64 10_Targets Over_dim_tied 32 10_Targets Over_dim_tied 256 Mnist Over_dim_tied 128 Mnist Over_dim_tied 64 Mnist Over_dim_tied 32 Mnist
0 0.9088 0.9152 0.9 0.9155 0.8287 0.8105 0.8194 0.8279
1 0.8869 0.8971 0.8595 0.8825 0.8578 0.8349 0.8613 0.8565
2 0.8847 0.8952 0.7574 0.8798 0.8558 0.8264 0.8609 0.8463
3 0.8844 0.8949 0.738 0.8792 0.8407 0.812 0.849 0.814
4 0.8844 0.8949 0.7051 0.879 0.8166 0.7912 0.8305 0.7814
5 0.8844 0.8949 0.7049 0.879 0.7911 0.7676 0.7998 0.7402
6 0.8844 0.8949 0.7046 0.879 0.7658 0.7422 0.7728 0.7004
7 0.8844 0.8949 0.7046 0.879 0.7377 0.7218 0.7449 0.6663
Loss over iterations autoencoder
Over_dim_tied 256 10_Targets Over_dim_tied 128 10_Targets Over_dim_tied 64 10_Targets Over_dim_tied 32 10_Targets Over_dim_tied 256 Mnist Over_dim_tied 128 Mnist Over_dim_tied 64 Mnist Over_dim_tied 32 Mnist
0 0.40483 0.402739 0.383918 0.396252 0.103948 2.52575e+06 7.13224e+06 0.0977971
1 0.425784 0.425326 0.414778 0.421516 425.388 2.79519e+06 5.77105e+21 0.106349
2 0.427908 40.7737 0.439398 0.426382 3164.44 2.80429e+06 inf 1.78619e+07
3 0.428039 2.63022e+12 0.445072 0.42703 9309.65 2.84269e+06 inf 6.47028e+22
4 0.42804 1.72183e+23 0.453312 0.427173 15798.3 2.91678e+06 inf inf
5 0.42804 inf 0.455866 0.42718 23702.4 2.95792e+06 nan inf
6 0.42804 inf 0.456654 0.42718 32169.6 3.10959e+06 nan inf
7 0.42804 inf 0.456781 0.42718 45090.3 3.27197e+06 nan nan
MAE over iterations autoencoder
Over_dim_tied 256 10_Targets Over_dim_tied 128 10_Targets Over_dim_tied 64 10_Targets Over_dim_tied 32 10_Targets Over_dim_tied 256 Mnist Over_dim_tied 128 Mnist Over_dim_tied 64 Mnist Over_dim_tied 32 Mnist
0 0.274157 0.2737 0.275191 0.27277 0.119765 218.981 224.646 0.123857
1 0.279011 0.278898 0.284809 0.279166 0.407303 235.488 6.38703e+09 0.125561
2 0.279475 0.334857 0.293963 0.280419 2.15254 235.99 1.81688e+17 38.3152
3 0.2795 14137.5 0.296832 0.280609 6.13899 239.226 5.16834e+24 2.29782e+09
4 0.2795 3.61709e+09 0.299741 0.280655 10.2659 245.573 1.4702e+32 1.38298e+17
5 0.2795 9.2546e+14 0.300467 0.280657 15.3625 249.071 nan 8.32372e+24
6 0.2795 2.36786e+20 0.300692 0.280657 20.8374 262.354 nan inf
7 0.2795 6.05835e+25 0.300727 0.280657 29.1036 276.142 nan nan 
predictions_df_50
Accuracy over iterations evaluations_feature_classifier
Over_dim_tied 256 10_Targets Over_dim_tied 128 10_Targets Over_dim_tied 64 10_Targets Over_dim_tied 32 10_Targets Over_dim_tied 256 Mnist Over_dim_tied 128 Mnist Over_dim_tied 64 Mnist Over_dim_tied 32 Mnist
0 0.8586 0.8754 0.8584 0.8791 0.7813 0.7421 0.7715 0.7569
1 0.8301 0.8483 0.8121 0.8437 0.8109 0.7673 0.8078 0.7876
2 0.8281 0.8462 0.714 0.8386 0.8035 0.7587 0.8115 0.7724
3 0.8278 0.846 0.6961 0.8376 0.7856 0.7406 0.7972 0.7485
4 0.8278 0.846 0.6701 0.8375 0.7588 0.7183 0.7785 0.7082
5 0.8278 0.846 0.6698 0.8375 0.7325 0.6929 0.7458 0.6732
6 0.8278 0.846 0.6697 0.8375 0.7005 0.6692 0.7165 0.6337
7 0.8278 0.846 0.6697 0.8375 0.6745 0.6449 0.6857 0.5984
Loss over iterations autoencoder
Over_dim_tied 256 10_Targets Over_dim_tied 128 10_Targets Over_dim_tied 64 10_Targets Over_dim_tied 32 10_Targets Over_dim_tied 256 Mnist Over_dim_tied 128 Mnist Over_dim_tied 64 Mnist Over_dim_tied 32 Mnist
0 0.410291 0.403392 0.380766 0.396454 0.12721 4.42122e+06 2.21598e+07 1850.55
1 0.438458 0.436516 0.419941 0.428222 211.097 4.81468e+06 1.7931e+22 6.70223e+18
2 0.441038 0.441116 0.4471 0.434723 5219.3 4.81468e+06 inf inf
3 0.441256 0.441503 0.453599 0.435611 13886.4 4.86555e+06 inf inf
4 0.44128 0.441517 0.461136 0.435791 22840.4 4.94813e+06 inf inf
5 0.44128 0.441517 0.463388 0.435797 35487.4 5.08258e+06 nan nan
6 0.44128 0.441517 0.46404 0.435797 47560.2 5.21522e+06 nan nan
7 0.44128 0.441517 0.464181 0.435797 61819.9 5.37646e+06 nan nan
MAE over iterations autoencoder
Over_dim_tied 256 10_Targets Over_dim_tied 128 10_Targets Over_dim_tied 64 10_Targets Over_dim_tied 32 10_Targets Over_dim_tied 256 Mnist Over_dim_tied 128 Mnist Over_dim_tied 64 Mnist Over_dim_tied 32 Mnist
0 0.279652 0.277233 0.277323 0.27522 0.135244 380.437 460.318 0.527041
1 0.286171 0.285365 0.289057 0.283317 0.270596 405.093 1.30907e+10 2.33864e+07
2 0.28676 0.286546 0.298645 0.285066 3.52272 405.097 3.72382e+17 1.40755e+15
3 0.28681 0.286628 0.301496 0.285337 9.10691 409.396 1.05929e+25 8.4716e+22
4 0.286814 0.286628 0.304095 0.285386 14.9043 416.579 3.01329e+32 5.09878e+30
5 0.286814 0.286628 0.304733 0.285387 22.9272 427.691 nan nan
6 0.286814 0.286628 0.304926 0.285387 30.7271 439.112 nan nan
7 0.286814 0.286628 0.304971 0.285387 40.0025 452.507 nan nan 
predictions_df_60
Accuracy over iterations evaluations_feature_classifier
Over_dim_tied 256 10_Targets Over_dim_tied 128 10_Targets Over_dim_tied 64 10_Targets Over_dim_tied 32 10_Targets Over_dim_tied 256 Mnist Over_dim_tied 128 Mnist Over_dim_tied 64 Mnist Over_dim_tied 32 Mnist
0 0.7873 0.8113 0.7951 0.8235 0.721 0.6804 0.7113 0.6732
1 0.7557 0.7794 0.7474 0.785 0.7473 0.702 0.7428 0.7074
2 0.7525 0.7757 0.6544 0.781 0.741 0.6928 0.7433 0.6904
3 0.7521 0.7755 0.6377 0.7804 0.7192 0.6732 0.732 0.6618
4 0.7521 0.7754 0.6169 0.7803 0.6905 0.6494 0.7127 0.6257
5 0.7521 0.7754 0.6163 0.7802 0.6663 0.6243 0.6778 0.5902
6 0.7521 0.7754 0.6163 0.7802 0.6406 0.5997 0.6528 0.5511
7 0.7521 0.7754 0.616 0.7802 0.6083 0.5752 0.6244 0.5171
Loss over iterations autoencoder
Over_dim_tied 256 10_Targets Over_dim_tied 128 10_Targets Over_dim_tied 64 10_Targets Over_dim_tied 32 10_Targets Over_dim_tied 256 Mnist Over_dim_tied 128 Mnist Over_dim_tied 64 Mnist Over_dim_tied 32 Mnist
0 0.417398 0.408326 0.379036 0.399529 0.152823 5.86798e+06 2.44959e+07 0.143526
1 0.453891 0.451729 0.428932 0.439389 1174.54 6.38967e+06 1.98213e+22 6.82105e+10
2 0.457474 0.457266 0.458879 0.446559 10499.5 6.4281e+06 inf 2.47088e+26
3 0.457815 0.457804 0.466308 0.447465 19437.8 6.47932e+06 inf inf
4 0.457817 0.457856 0.472902 0.4476 33271.3 6.53314e+06 inf inf
5 0.457817 0.457878 0.474797 0.447651 45289.9 6.64166e+06 nan inf
6 0.457817 0.457883 0.475353 0.447651 57509 6.85152e+06 nan nan
7 0.457817 0.457883 0.475509 0.447651 69673.2 7.09927e+06 nan nan
MAE over iterations autoencoder
Over_dim_tied 256 10_Targets Over_dim_tied 128 10_Targets Over_dim_tied 64 10_Targets Over_dim_tied 32 10_Targets Over_dim_tied 256 Mnist Over_dim_tied 128 Mnist Over_dim_tied 64 Mnist Over_dim_tied 32 Mnist
0 0.286747 0.283158 0.281009 0.279812 0.150939 504.641 507.256 0.155433
1 0.295163 0.293912 0.295556 0.289611 0.933239 537.58 1.44256e+10 2754.45
2 0.295991 0.295308 0.305509 0.291539 6.94743 540.821 4.10354e+17 1.65767e+11
3 0.296073 0.295427 0.308378 0.2918 12.6422 545.13 1.16731e+25 9.97699e+18
4 0.296073 0.295442 0.310582 0.291845 21.6395 549.941 3.32056e+32 6.00483e+26
5 0.296073 0.295451 0.311118 0.291862 29.2726 559.337 nan inf
6 0.296073 0.295452 0.311281 0.291862 37.1904 577.29 nan nan
7 0.296073 0.295452 0.311336 0.291862 44.9696 598.505 nan nan 
predictions_df_70
Accuracy over iterations evaluations_feature_classifier
Over_dim_tied 256 10_Targets Over_dim_tied 128 10_Targets Over_dim_tied 64 10_Targets Over_dim_tied 32 10_Targets Over_dim_tied 256 Mnist Over_dim_tied 128 Mnist Over_dim_tied 64 Mnist Over_dim_tied 32 Mnist
0 0.712 0.7241 0.7228 0.7543 0.6585 0.6003 0.634 0.5814
1 0.6755 0.6837 0.6686 0.7148 0.6754 0.6194 0.6654 0.6096
2 0.6714 0.68 0.5845 0.71 0.6622 0.6056 0.6668 0.5922
3 0.671 0.6799 0.5691 0.7094 0.6444 0.5839 0.6567 0.5652
4 0.671 0.68 0.554 0.7093 0.6227 0.5642 0.6363 0.5351
5 0.671 0.68 0.5537 0.7093 0.5987 0.5458 0.6048 0.5058
6 0.671 0.68 0.5536 0.7093 0.5682 0.5237 0.5809 0.4762
7 0.671 0.68 0.5536 0.7093 0.5425 0.5001 0.557 0.4506
Loss over iterations autoencoder
Over_dim_tied 256 10_Targets Over_dim_tied 128 10_Targets Over_dim_tied 64 10_Targets Over_dim_tied 32 10_Targets Over_dim_tied 256 Mnist Over_dim_tied 128 Mnist Over_dim_tied 64 Mnist Over_dim_tied 32 Mnist
0 0.424893 0.416433 0.379668 0.40174 0.179462 7.88446e+06 4.20404e+07 51443.5
1 0.469254 0.471375 0.441857 0.449146 1894.19 8.42567e+06 3.40178e+22 1.8635e+20
2 0.473659 0.478921 0.474173 0.458057 11057.2 8.45348e+06 inf inf
3 0.474073 0.479675 0.482592 0.459267 23399.3 8.53513e+06 inf inf
4 0.474074 0.479719 0.488431 0.459431 34722.8 8.59214e+06 inf inf
5 0.474074 0.479723 0.490332 0.459438 47389 8.70958e+06 nan nan
6 0.474074 0.479722 0.490855 0.459439 62673.1 8.87507e+06 nan nan
7 0.474074 0.479722 0.49099 0.459439 75598.8 9.00403e+06 nan nan
MAE over iterations autoencoder
Over_dim_tied 256 10_Targets Over_dim_tied 128 10_Targets Over_dim_tied 64 10_Targets Over_dim_tied 32 10_Targets Over_dim_tied 256 Mnist Over_dim_tied 128 Mnist Over_dim_tied 64 Mnist Over_dim_tied 32 Mnist
0 0.293888 0.290865 0.28649 0.284382 0.166536 675.241 766.394 2.2191
1 0.30403 0.305018 0.304186 0.295676 1.40041 708.841 2.17967e+10 1.2332e+08
2 0.305074 0.306905 0.314251 0.297984 7.31724 711.446 6.20036e+17 7.42222e+15
3 0.305172 0.307086 0.317167 0.298324 15.3404 718.284 1.76378e+25 4.46719e+23
4 0.305173 0.307094 0.319062 0.29837 22.5776 722.886 5.0173e+32 2.68866e+31
5 0.305173 0.307092 0.319614 0.298372 30.6874 733.065 nan nan
6 0.305173 0.307091 0.319766 0.298372 40.4393 746.899 nan nan
7 0.305173 0.307091 0.319809 0.298372 48.7669 757.779 nan nan 
predictions_df_80
Accuracy over iterations evaluations_feature_classifier
Over_dim_tied 256 10_Targets Over_dim_tied 128 10_Targets Over_dim_tied 64 10_Targets Over_dim_tied 32 10_Targets Over_dim_tied 256 Mnist Over_dim_tied 128 Mnist Over_dim_tied 64 Mnist Over_dim_tied 32 Mnist
0 0.6201 0.6322 0.637 0.6695 0.5815 0.5269 0.5545 0.4979
1 0.5872 0.5943 0.5879 0.6334 0.5949 0.5324 0.5839 0.5232
2 0.583 0.5903 0.5121 0.631 0.5837 0.5207 0.5849 0.5108
3 0.5827 0.5902 0.5008 0.6298 0.5653 0.5084 0.5762 0.4801
4 0.5827 0.5902 0.4935 0.6298 0.54 0.4923 0.5609 0.456
5 0.5827 0.5902 0.4933 0.6297 0.5183 0.4721 0.5341 0.4309
6 0.5827 0.5902 0.493 0.6297 0.4926 0.4559 0.5161 0.413
7 0.5827 0.5902 0.493 0.6297 0.4689 0.443 0.4974 0.3923
Loss over iterations autoencoder
Over_dim_tied 256 10_Targets Over_dim_tied 128 10_Targets Over_dim_tied 64 10_Targets Over_dim_tied 32 10_Targets Over_dim_tied 256 Mnist Over_dim_tied 128 Mnist Over_dim_tied 64 Mnist Over_dim_tied 32 Mnist
0 0.436443 0.430405 0.384682 0.406368 0.208251 9.74844e+06 5.6217e+07 132070
1 0.486314 1.09131e+08 0.454558 0.462524 3732.36 1.04488e+07 4.54892e+22 4.78392e+20
2 0.490589 7.14404e+18 0.488003 0.471316 14617.4 1.04615e+07 inf inf
3 0.490844 4.67671e+29 0.496666 0.472457 27537.2 1.04881e+07 inf inf
4 0.490855 inf 0.501108 0.472657 39231 1.06762e+07 inf inf
5 0.490855 inf 0.502375 0.472676 50583.6 1.08189e+07 nan nan
6 0.490855 inf 0.502751 0.472676 64751.5 1.09616e+07 nan nan
7 0.490855 inf 0.502837 0.472676 81520.3 1.11892e+07 nan nan
MAE over iterations autoencoder
Over_dim_tied 256 10_Targets Over_dim_tied 128 10_Targets Over_dim_tied 64 10_Targets Over_dim_tied 32 10_Targets Over_dim_tied 256 Mnist Over_dim_tied 128 Mnist Over_dim_tied 64 Mnist Over_dim_tied 32 Mnist
0 0.302497 0.300678 0.293671 0.290125 0.182696 834.933 935.025 3.90084
1 0.313446 91.379 0.31238 0.303022 2.59676 879.023 2.65933e+10 2.23655e+08
2 0.314394 2.3299e+07 0.322027 0.305147 9.62583 880.099 7.56482e+17 1.34611e+16
3 0.314446 5.96122e+12 0.32468 0.305444 17.8886 882.532 2.15191e+25 8.10179e+23
4 0.314448 1.52522e+18 0.326063 0.305497 25.3939 898.28 6.1214e+32 4.8762e+31
5 0.314449 3.9024e+23 0.326419 0.305501 32.6098 910.452 nan nan
6 0.314449 9.98458e+28 0.326535 0.305501 41.8009 922.794 nan nan
7 0.314449 inf 0.326563 0.305501 52.5019 941.838 nan nan 
predictions_df_90
Accuracy over iterations evaluations_feature_classifier
Over_dim_tied 256 10_Targets Over_dim_tied 128 10_Targets Over_dim_tied 64 10_Targets Over_dim_tied 32 10_Targets Over_dim_tied 256 Mnist Over_dim_tied 128 Mnist Over_dim_tied 64 Mnist Over_dim_tied 32 Mnist
0 0.5258 0.5347 0.5477 0.5696 0.4877 0.4276 0.4648 0.4185
1 0.4909 0.4943 0.5 0.5349 0.5019 0.4427 0.494 0.4364
2 0.4878 0.4904 0.4366 0.5314 0.493 0.4332 0.4943 0.426
3 0.4879 0.4901 0.4272 0.5308 0.48 0.4205 0.4836 0.4056
4 0.4879 0.4902 0.4216 0.5307 0.4597 0.4068 0.467 0.3886
5 0.4879 0.4902 0.4212 0.5307 0.444 0.3921 0.44 0.3664
6 0.4879 0.4902 0.4212 0.5307 0.4305 0.379 0.4254 0.3489
7 0.4879 0.4902 0.4212 0.5307 0.4108 0.3668 0.4126 0.3301
Loss over iterations autoencoder
Over_dim_tied 256 10_Targets Over_dim_tied 128 10_Targets Over_dim_tied 64 10_Targets Over_dim_tied 32 10_Targets Over_dim_tied 256 Mnist Over_dim_tied 128 Mnist Over_dim_tied 64 Mnist Over_dim_tied 32 Mnist
0 0.445065 0.437497 0.387168 0.41089 0.240506 1.35461e+07 6.57491e+07 352749
1 0.502291 0.51108 0.467161 0.475999 2534.91 1.47513e+07 5.32023e+22 1.27771e+21
2 0.506685 0.519687 0.50146 0.485716 15143.6 1.47717e+07 inf inf
3 0.507073 0.520529 0.511031 0.486989 29559.8 1.4879e+07 inf inf
4 0.507081 0.520673 0.515045 0.487153 42813.9 1.49731e+07 inf inf
5 0.507081 0.520686 0.516242 0.487179 56596.2 1.51798e+07 nan nan
6 0.507081 0.520686 0.516614 0.48718 69377.1 1.53626e+07 nan nan
7 0.507081 0.520686 0.516744 0.48718 83560.9 1.55018e+07 nan nan
MAE over iterations autoencoder
Over_dim_tied 256 10_Targets Over_dim_tied 128 10_Targets Over_dim_tied 64 10_Targets Over_dim_tied 32 10_Targets Over_dim_tied 256 Mnist Over_dim_tied 128 Mnist Over_dim_tied 64 Mnist Over_dim_tied 32 Mnist
0 0.310092 0.308183 0.299873 0.295911 0.199425 1164.99 1117.59 6.92175
1 0.322415 0.326559 0.320382 0.31062 1.84678 1240.91 3.17863e+10 4.04333e+08
2 0.323316 0.328659 0.329699 0.312928 9.97175 1243 9.04203e+17 2.43355e+16
3 0.3234 0.328872 0.332457 0.313238 19.214 1251.68 2.57212e+25 1.46467e+24
4 0.3234 0.328912 0.333667 0.313284 27.7209 1260.11 7.31675e+32 inf
5 0.3234 0.328915 0.333999 0.313288 36.5904 1277.21 nan nan
6 0.3234 0.328915 0.33411 0.313289 44.7494 1292.49 nan nan
7 0.3234 0.328915 0.334153 0.313289 53.8455 1304.25 nan nan 
predictions_df_100
Accuracy over iterations evaluations_feature_classifier
Over_dim_tied 256 10_Targets Over_dim_tied 128 10_Targets Over_dim_tied 64 10_Targets Over_dim_tied 32 10_Targets Over_dim_tied 256 Mnist Over_dim_tied 128 Mnist Over_dim_tied 64 Mnist Over_dim_tied 32 Mnist
0 0.4324 0.4416 0.4537 0.4836 0.4127 0.3567 0.3812 0.3548
1 0.4006 0.4066 0.4127 0.4543 0.4223 0.3616 0.4079 0.3614
2 0.3986 0.4038 0.3644 0.452 0.4144 0.3535 0.4062 0.3526
3 0.3983 0.403 0.3585 0.4519 0.4014 0.3467 0.3978 0.3373
4 0.3982 0.403 0.3548 0.4518 0.3811 0.3335 0.3897 0.3232
5 0.3982 0.403 0.3545 0.4518 0.3662 0.3195 0.3625 0.3048
6 0.3982 0.403 0.3542 0.4518 0.3526 0.3121 0.3544 0.2909
7 0.3982 0.403 0.3542 0.4518 0.338 0.3016 0.3454 0.2828
Loss over iterations autoencoder
Over_dim_tied 256 10_Targets Over_dim_tied 128 10_Targets Over_dim_tied 64 10_Targets Over_dim_tied 32 10_Targets Over_dim_tied 256 Mnist Over_dim_tied 128 Mnist Over_dim_tied 64 Mnist Over_dim_tied 32 Mnist
0 0.454725 0.45108 0.39497 0.419231 0.27189 1.89169e+07 9.15463e+07 7.21208e+06
1 0.518014 0.53049 0.481865 0.489098 2953.55 2.03214e+07 7.40765e+22 2.61243e+22
2 0.523514 0.539329 0.516644 0.498944 14789.5 2.03342e+07 inf inf
3 0.523904 0.54019 0.526589 0.499971 31879 2.03836e+07 inf inf
4 0.523926 0.540274 0.530323 0.500032 46811.5 2.04392e+07 inf inf
5 0.523928 0.540276 0.531252 0.500087 62903.4 2.06526e+07 nan nan
6 0.523928 0.540276 0.531524 0.500093 77770.4 2.08556e+07 nan nan
7 0.523928 0.540276 0.531578 0.500093 94271 2.10574e+07 nan nan
MAE over iterations autoencoder
Over_dim_tied 256 10_Targets Over_dim_tied 128 10_Targets Over_dim_tied 64 10_Targets Over_dim_tied 32 10_Targets Over_dim_tied 256 Mnist Over_dim_tied 128 Mnist Over_dim_tied 64 Mnist Over_dim_tied 32 Mnist
0 0.318187 0.317664 0.307869 0.302561 0.21702 1620.48 1661.19 63.6357
1 0.331444 0.337099 0.329051 0.317666 2.11387 1709.43 4.72491e+10 3.81653e+09
2 0.332533 0.339079 0.337847 0.319945 9.787 1710.51 1.34406e+18 2.29704e+17
3 0.332607 0.339285 0.340557 0.320195 20.8379 1714.74 3.82337e+25 1.38252e+25
4 0.332612 0.339305 0.341615 0.32021 30.3468 1719.8 1.08761e+33 inf
5 0.332612 0.339305 0.341869 0.320231 40.5508 1737.37 nan nan
6 0.332612 0.339305 0.341938 0.320233 50.1415 1754.52 nan nan
7 0.332612 0.339305 0.341949 0.320233 60.8013 1771.73 nan nan